Automatic Questioning and Answering Processing Method and Automatic Questioning and Answering System

1. An automatic question-answer (QA) processing method performed at a computing device having one or more processors and memory storing a plurality of programs to be executed by the one or more processors, the method comprising: obtaining, after receiving a target question text, a target keyword corresponding to the target question text; determining candidate question texts that match the target keyword by (i) identifying question texts in a pre-collected QA data pair using a pre-established inverse index relationship between the target keyword and the question texts, each question text having an associated answer and a weight value corresponding to the target keyword, and (ii) selecting top N of the question texts according to their associated weight values as the candidate question texts, that match the target keyword, wherein N is greater than 1; calculating a semantic similarity value between each of the candidate question texts and the target question text; determining, from the pre-collected QA data pair, an answer associated with one of the candidate question texts corresponding to the target question text having a highest semantic similarity value, and before obtaining a target keyword corresponding to the target question text: obtaining a question text from the pre-collected QA data pair; obtaining a keyword corresponding to the question text; and establishing an index relationship between the keyword and the question text; and the operation of determining a candidate question text that is in a pre-collected QA data pair and that matches the target keyword comprises: determining, in the established index relationship between the keyword and the question text, a first keyword that matches the target keyword, obtaining a question text that has an index relationship with the first keyword, and using the question text as a candidate question text of the target question text.

2. The automatic QA processing method according to claim 1 , wherein the operation of obtaining a target keyword corresponding to the target question text comprises: performing word segmentation processing on the target question text, to obtain a target candidate word in the target question text; calculating a weight value separately corresponding to each target candidate word in the target question text; and determining first M candidate words with largest weight values as the target keyword corresponding to the target question text, wherein M is a natural number.

3. The automatic QA processing method according to claim 2 , further comprising: before calculating a weight value separately corresponding to each target candidate word in the target question text: screening out a word of a preset type in the target candidate word.

4. The automatic QA processing method according to claim 1 , wherein the operation of obtaining a keyword corresponding to the question text comprises: performing word segmentation processing on the question text, to obtain a candidate word in the question text; calculating a weight value separately corresponding to each candidate word in the question text; and determining first N candidate words with largest weight values as the keyword corresponding to the question text, wherein N is a natural number.

5. The automatic QA processing method according to claim 1 , wherein the operation of calculating a semantic similarity value between each candidate question text and the target question text comprises: determining, by using a pre-trained word embedding model, semantic vectors respectively corresponding to each candidate question text and the target question text; and calculating, for each candidate question text, a vector distance between the semantic vector corresponding to the candidate question text and the semantic vector corresponding to the target question text, and using the vector distance as the semantic similarity value between the candidate question text and the target question text.

6. The automatic QA processing method according to claim 5 , wherein the operation of determining, based on the semantic similarity value, an answer corresponding to the target question text comprises: using, as the answer corresponding to the target question text, an answer corresponding to a candidate question text corresponding to a maximum semantic similarity value in the pre-collected QA data pair.

7. The automatic QA processing method according to claim 1 , further comprising: calculating, for each candidate question text, an editing distance between the candidate question text and the target question text, and using the editing distance as a character string similarity value between the candidate question text and the target question text; and the operation of determining, based on the semantic similarity value, an answer corresponding to the target question text comprises: determining, with reference to the semantic similarity value and the character string similarity value between the candidate question text and the target question text, the answer corresponding to the target question text.

8. An automatic question-answer (QA) system, comprising: one or more processors; memory coupled to the one or more processors; and a plurality of computer-readable instructions that, when executed by the one or more processors, cause the server to perform the following operations: obtaining, after receiving a target question text, a target keyword corresponding to the target question text; determining candidate question texts that match the target keyword by (i) identifying question texts in a pre-collected QA data pair using a pre-established inverse index relationship between the target keyword and the question texts, each question text having an associated answer and a weight value corresponding to the target keyword, and (ii) selecting top N of the question texts according to their associated weight values as the candidate question texts, that match the target keyword, wherein N is greater than 1; calculating a semantic similarity value between each of the candidate question texts and the target question text; determining, from the pre-collected QA data pair, an answer associated with one of the candidate question texts corresponding to the target question text having a highest semantic similarity value; before obtaining a target keyword corresponding to the target question text: obtaining a question text from the pre-collected QA data pair; obtaining a keyword corresponding to the question text; and establishing an index relationship between the keyword and the question text; and the operation of determining a candidate question text that is in a pre-collected QA data pair and that matches the target keyword comprises: determining, in the established index relationship between the keyword and the question text, a first keyword that matches the target keyword, obtaining a question text that has an index relationship with the first keyword, and using the question text as a candidate question text of the target question text.

9. The automatic QA system according to claim 8 , wherein the operation of obtaining a target keyword corresponding to the target question text comprises: performing word segmentation processing on the target question text, to obtain a target candidate word in the target question text; calculating a weight value separately corresponding to each target candidate word in the target question text; and determining first M candidate words with largest weight values as the target keyword corresponding to the target question text, wherein M is a natural number.

10. The automatic QA system according to claim 9 , wherein the operations further comprise: before calculating a weight value separately corresponding to each target candidate word in the target question text: screening out a word of a preset type in the target candidate word.

11. The automatic QA system according to claim 8 , wherein the operation of obtaining a keyword corresponding to the question text comprises: performing word segmentation processing on the question text, to obtain a candidate word in the question text; calculating a weight value separately corresponding to each candidate word in the question text; and determining first N candidate words with largest weight values as the keyword corresponding to the question text, wherein N is a natural number.

12. The automatic QA system according to claim 8 , wherein the operations further comprise: wherein the operation of calculating a semantic similarity value between each candidate question text and the target question text comprises: determining, by using a pre-trained word embedding model, semantic vectors respectively corresponding to each candidate question text and the target question text; and calculating, for each candidate question text, a vector distance between the semantic vector corresponding to the candidate question text and the semantic vector corresponding to the target question text, and using the vector distance as the semantic similarity value between the candidate question text and the target question text.

13. The automatic QA system according to claim 12 , wherein the operation of determining, based on the semantic similarity value, an answer corresponding to the target question text comprises: using, as the answer corresponding to the target question text, an answer corresponding to a candidate question text corresponding to a maximum semantic similarity value in the pre-collected QA data pair.

14. The automatic QA system according to claim 8 , wherein the operations further comprise: calculating, for each candidate question text, an editing distance between the candidate question text and the target question text, and using the editing distance as a character string similarity value between the candidate question text and the target question text; and the operation of determining, based on the semantic similarity value, an answer corresponding to the target question text comprises: determining, with reference to the semantic similarity value and the character string similarity value between the candidate question text and the target question text, the answer corresponding to the target question text.

15. A non-transitory computer readable storage medium storing a plurality of instructions in connection with a computing device having one or more processors, wherein the plurality of instructions, when executed by the one or more processors, cause the server to perform a plurality of operations including: obtaining, after receiving a target question text, a target keyword corresponding to the target question text; determining candidate question texts that match the target keyword by (i) identifying question texts in a pre-collected QA data pair using a pre-established inverse index relationship between the target keyword and the question texts, each question text having an associated answer and a weight value corresponding to the target keyword, and (ii) selecting top N of the question texts according to their associated weight values as the candidate question texts, that match the target keyword, wherein N is greater than 1; calculating a semantic similarity value between each of the candidate question texts and the target question text; determining, from the pre-collected QA data pair, an answer associated with one of the candidate question texts corresponding to the target question text having a highest semantic similarity value, before obtaining a target keyword corresponding to the target question text: obtaining a question text from the pre-collected QA data pair; obtaining a keyword corresponding to the question text; and establishing an index relationship between the keyword and the question text; and the operation of determining a candidate question text that is in a pre-collected QA data pair and that matches the target keyword comprises: determining, in the established index relationship between the keyword and the question text, a first keyword that matches the target keyword, obtaining a question text that has an index relationship with the first keyword, and using the question text as a candidate question text of the target question text.

16. The non-transitory computer readable storage medium according to claim 15 , wherein the operation of obtaining a target keyword corresponding to the target question text comprises: performing word segmentation processing on the target question text, to obtain a target candidate word in the target question text; calculating a weight value separately corresponding to each target candidate word in the target question text; and determining first M candidate words with largest weight values as the target keyword corresponding to the target question text, wherein M is a natural number.

17. The non-transitory computer readable storage medium according to claim 15 , wherein the operations further comprise: wherein the operation of calculating a semantic similarity value between each candidate question text and the target question text comprises: determining, by using a pre-trained word embedding model, semantic vectors respectively corresponding to each candidate question text and the target question text; and calculating, for each candidate question text, a vector distance between the semantic vector corresponding to the candidate question text and the semantic vector corresponding to the target question text, and using the vector distance as the semantic similarity value between the candidate question text and the target question text.